# Hot metal desulphurisation

Hot metal desulphurisation The hot metal produced with blast furnaces contains relatively high amounts of sulphur, which is detrimental for the mechanical properties of steel. Owing to the limited desulphurisation capacity of converter processes, it is commonplace to conduct hot metal desulphurisation in a ladle or a torpedo car before further treatment at the meltshop. The desulphurisation reagent is typically injected through a submedged top lance.## Mathematical modelling

Our recent research is related to development of a mathematical model for hot metal desulphurisation. The objective is to derive a model, which accounts for the effects of main technological and operational parameters on the desulphurisation efficiency. The model calculates the desulphurisation rate based on the basis of thermodynamic driving force and relevant mass transfer resistances. A detailed description of the model will be published in the near future.## Data-driven modelling

Alongside the phenomenon-based
approach, data-driven methods (e.g. GA and ANN) have been applied for
predicting hot metal desulphurisation as well as sulphide capacity.
This model employs genetic algorithm for determining the values of the
model parameters based on plant data.## Numerical and physical modelling of fluid flows

Numerical and phyisical modelling of fluid flows aims to establish the fluid flow field during reagent injection. In this way, it is possible to gather new information for optimisation of the injection practice, e.g. reagent size distribution, type of lance and carrier gas flow rate.## High-temperature experiments

In addition mathematical modelling, high temperature experiments are carried out to investigate kinetics of metal-slag reactions during and after the hot metal desulphurisation.## Results

### Highlights

- A mathematical model was developed for hot metal desulphurisation. The model can be employed for studying effect of the main operating and technological parameters (e.g. reagent composition, size distribution and injection rate) on the efficiency of desulphurisation.
- Data-driven parametrised reaction model, in which the rate parameters are optimised with an genetic algorithm.
- A genetic algorithm based variable selection method was developed for prediction of hot metal desulphurisation

### Publications

- T. Vuolio, V.-V. Visuri, T. Paananen, and T. Fabritius,
“Identification of rate, extent and mechanisms of hot metal
resulfurization with CaO–SiO
_{2}–NaO_{2}slag systems,” Metallurgical and Materials Transactions B, forthcoming.

- V.-V. Visuri, P. Sulasalmi, T. Vuolio, T. Paananen, T. Haas, H. Pfeifer, and T. Fabritius, "Mathematical Modelling of the Effect of Reagent Particle Size Distribution on the Efficiency of Hot Metal Desulphurisation", Proceedings of the 4th European Steel Technology and Application Days, Steel Institute VDEh, Düsseldorf, Germany, 2019, forthcoming.
- T. Vuolio, V.-V. Visuri, A. Sorsa, T. Paananen, and T. Fabritius, "Genetic Algorithm Based Variable Selection in Prediction of Hot Metal Desulfurization Kinetics", Steel Research International, forthcoming.
- T. Vuolio, V.-V. Visuri, S. Tuomikoski, T. Paananen, and T. Fabritius, "Data-Driven Mathematical Modeling of the Effect of Particle Size Distribution on the Transitory Reaction Kinetics of Hot Metal Desulfurization", Metallurgical and Materials Transactions B, vol. 49, no. 5, pp. 2692–2708, 2018.

### Theses instructed

- P. Pekuri, "Effect of initial slag on the efficiency of hot metal desulphurisation", in progress.
- P. Lehtonen, "Experimental Investigation of Hot Metal Desulphurisation", Master's thesis, University of Oulu, 2017.
- T. Vuolio, "Improvement potential of primary hot metal delsulphurization", Master's thesis, University of Oulu, 2017.